Metformin Use and Thirty-Day Readmission Among Patients with Bipolar Disorders: A Propensity Score-matching Analysis

Metformin may affect patients with multi-comorbid bipolar disorder through the integrated stress response. However, the relationship between them remains unclear. This study aimed to explore the relationship between metformin use and early (30-day) readmission in bipolar disorder patients. Data were extracted from the Medical Information Mart for Intensive Care IV database. Adult patients with a documented diagnosis of bipolar disorder were screened. Multivariable logistic regression and propensity score matching were used to investigate potential associations. Data on 1688 patients were included. The crude early (30-day) readmission was significantly higher in patients with metformin use (17/114 vs. 132/1574, p = 0.033). In the extended multivariable logistic models, the hazard ratio (HR) of metformin use was consistently significant in all models (HR range 1.87–2.99, p < 0.05 for all). After propensity score matching, the early (30-day) readmission remained significantly higher in the metformin use group (Mahalanobis matching, |t|>1.96). Although residual confounding cannot be excluded, metformin use is associated with higher readmission.